What Do You Do When AI Beats You?

When Beethoven was at the peak of his career, several of his contemporaries struggled to deal with the realization that they may never create anything that lived up to his creations. Brahms, for example, refused to make a symphony for 21 years. Schubert is quoted as saying, "Who can ever do anything after Beethoven?"

We're apparently seeing the same effect via Artificial Intelligence. 

When it comes to popular AI, not much surpasses the popularity of AI's growing chokehold on gaming. Recently, I've shared about AI winning at video games,  but in 2016 I shared about humans losing to AI in Go for the first time. 

What it is:  A computer has just beaten the world's best Go player. AlphaGo, a program created by Google-owned AI company DeepMind, beat European Go champ Fan Hui all five times they played in tournament conditions, and also won 99.8% of Go games against other computer programs. Unlike IBM's Deep Blue, which defeated chess champ Garry Kasparov, AlphaGo wasn't programmed to play Go. Instead, as Nature reports, it learned how to play via a general-purpose algorithm that interpreted the game's patterns.

Why it's important:  AlphaGo's learning technique means it can recognize complex patterns, long-term planning and decision-making: refined skills that were once stricly human in nature. Imagine the possibilities when neural networks like AlphaGo, infinite computing and the 'Internet of Everything' converge.

via - Google AI Algorithm Masters Ancient Game Of Go - February 2016

This week, a former Go champion who was beaten by DeepMind retired after "declaring AI invincible." 

Lee Se-dol quit for a couple of reasons. According to him, even if he's the #1 human, there's an undefeatable entity above him and he felt he had failed his country by losing to the AI. It's an unfair fight - AI plays untold millions of games to learn to play better and it doesn't get tired, bored, sick, distracted.  - we can't do that.

Much like Beethoven, AI is discouraging competition. 

Was Lee wrong to quit? It's hard to say, but as AI gets better at more activities, it's an issue we're going to see more often. There's always someone (or something) better - and a purely utilitarian approach isn't necessary or productive. 

I'm an advocate of intelligently adopting AI, and a believer that the scale of AI's "wins" is going to skyrocket - but I'm also a believer in the idea that "the game isn't over until I win". If I enjoy something, I'm not going to let 2^nd place stop me. 

 

 

The passionate pursuit of a goal is valuable regardless of the result, and bettering yourself at a skill - like Go - may not be a sustainable job, but it can still make a great hobby. 

AI is coming - but it doesn't have to be joy-sucking. 

Onwards!